Skip to content

Dataset · W8Yi

TCGA histological image embeddings (UNI2-h)

W8Yi/tcga-wsi-uni2h-features

Extracted feature representations of TCGA histological images for computational pathology research.

What's inside

Contains embeddings of whole slide image (WSI) patches from the TCGA oncology database extracted using the UNI2-h model, along with overlay images for quality control. Features fewer than a thousand records, organized by projects such as TCGA-HNSC and saved in H5 format. Descriptions are provided in English.

What you can do with it

  • Training classifiers and histological slide analysis models
  • Conducting computational pathology research without the overhead of extracting features from heavy WSI files
Tasks
Image feature extraction
Languages
English
Size
up to 1K rows
Downloads this month
20.1K
Likes
15

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("W8Yi/tcga-wsi-uni2h-features")